The etiology of prostate cancer is unknown, but there is compelling evidence that androgen, acting via the androgen receptor (AR), is required for the development of prostate cancer. AR gene mutations occur in only a subset of early stage prostate cancers. Therefore, if androgen is involved in prostate cancer etiology, it must primarily involve the normal, wild- type AR. The N-terminal transactivation domain of the AR contains a polymorphic glutamine repeat and a polymorphic glycine repeat; hence there are multiple normal AR alleles (variants). Our hypothesis is that polymorphic variants of the AR differ in activity and thereby differ in their ability to mediate the role of androgen in the development, progression, or growth of prostate cancer. We will use two approaches to test this hypothesis. One approach (Aim 1) will be to analyze AR alleles (glutamine and glycine repeat length in the AR gene) in men with and without prostate cancer to determine whether allele frequencies differ according to the presence or phenotype (classified according to race, stage, grade, and aggressiveness) of prostate cancer. The second approach (Aims 2 and 3) will be to determine how AR variants may affect androgen action. We will determine whether polymorphic AR variants differ in transcriptional activity and whether this is due to an effect of glutamine and/or glycine repeat length on intrinsic activity or on the affinity for proteins that interact with the AR and are required to transduce the androgenic signal. In Aim 2, the effect of glutamine and glycine repeat length on AR function will be tested by constructing AR expression vectors that contain glutamine and glycine repeat lengths that correspond to those that occur with the highest and lowest frequencies in men with prostate cancer; androgen binding and transcriptional activity of these AR variants will be measured by transfecting AR negative prostate cancer cells. Aim 3 is to determine the effect of glutamine and glycine repeat length on the interaction of the AR with its target proteins. We will identify proteins that interact with the glutamine and glycine repeat of the AR, and determine whether glutamine and/or glycine repeat length affect the strength of the interaction between the AR and this protein(s). The identification of AR-interacting targets should provide new insight into the mechanism by which androgen plays a critical role in prostate cancer etiology, and the mechanism by which AR polymorphisms may modify this role and thereby affect risk.